From: Lu Baolu <baolu.lu@linux.intel.com>
To: Ingo Molnar <mingo@kernel.org>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
Mathias Nyman <mathias.nyman@linux.intel.com>,
Ingo Molnar <mingo@redhat.com>,
tglx@linutronix.de, peterz@infradead.org,
linux-usb@vger.kernel.org, x86@kernel.org,
linux-kernel@vger.kernel.org
Subject: Re: [PATCH v5 1/4] usb: dbc: early driver for xhci debug capability
Date: Tue, 24 Jan 2017 12:44:39 +0800 [thread overview]
Message-ID: <5886DBB7.4070501@linux.intel.com> (raw)
In-Reply-To: <20170122090423.GA15061@gmail.com>
[-- Attachment #1: Type: text/plain, Size: 1954 bytes --]
Hi Ingo,
On 01/22/2017 05:04 PM, Ingo Molnar wrote:
> * Lu Baolu <baolu.lu@linux.intel.com> wrote:
>
>>>> +static void xdbc_runtime_delay(unsigned long count)
>>>> +{
>>>> + udelay(count);
>>>> +}
>>>> +static void (*xdbc_delay)(unsigned long) = xdbc_early_delay;
>>> Is this udelay() complication really necessary? udelay() should work fine even in
>>> early code. It might not be precisely calibrated, but should be good enough.
>> I tried udelay() in the early code. It's not precise enough for the
>> hardware handshaking.
> Possibly because on x86 early udelay() did not work at all - i.e. there's no delay
> whatsoever.
Yes.
>
> Could you try it on top of this commit in tip:timers/core:
>
> 4c45c5167c95 x86/timer: Make delay() work during early bootup
>
> ?
I tried tip:timers/core. It's not precise enough for my context either.
__const_udelay().
157 inline void __const_udelay(unsigned long xloops)
158 {
159 unsigned long lpj = this_cpu_read(cpu_info.loops_per_jiffy) ? : loops_per_jiffy;
160 int d0;
161
162 xloops *= 4;
163 asm("mull %%edx"
164 :"=d" (xloops), "=&a" (d0)
165 :"1" (xloops), "0" (lpj * (HZ / 4)));
166
167 __delay(++xloops);
168 }
In my early code, loops_per_jiffy is not initialized yet. Hence "lpj" for the asm line
is 4096 (default value).
The cpu_info.loops_per_jiffy actually reads 8832000 after initialization. They are
about 2000 times different.
I did a hacky test in kernel to check the difference between these two different
"lpj" values. (The hacky patch is attached.) Below is the output for 100ms delay.
[ 2.494751] udelay_test uninitialized ---->start
[ 2.494820] udelay_test uninitialized ---->end
[ 2.494828] udelay_test initialized ---->start
[ 2.595234] udelay_test initialized ---->end
For 100ms delay, udelay() with uninitialized loops_per_jiffy only gives a delay of
only 69us.
Best regards,
Lu Baolu
[-- Attachment #2: kernel-hack.patch --]
[-- Type: text/x-patch, Size: 2325 bytes --]
diff --git a/arch/x86/lib/delay.c b/arch/x86/lib/delay.c
index a8e91ae..ffc2874 100644
--- a/arch/x86/lib/delay.c
+++ b/arch/x86/lib/delay.c
@@ -168,6 +168,36 @@ inline void __const_udelay(unsigned long xloops)
}
EXPORT_SYMBOL(__const_udelay);
+void udelay_uninitialized(unsigned long xloops)
+{
+ unsigned long lpj = (1<<12);
+ int d0;
+
+ xloops *= 0x10c7ul;
+ xloops *= 4;
+ asm("mull %%edx"
+ :"=d" (xloops), "=&a" (d0)
+ :"1" (xloops), "0" (lpj * (HZ / 4)));
+
+ delay_loop(++xloops);
+}
+EXPORT_SYMBOL(udelay_uninitialized);
+
+void udelay_initialized(unsigned long xloops)
+{
+ unsigned long lpj = this_cpu_read(cpu_info.loops_per_jiffy);
+ int d0;
+
+ xloops *= 0x10c7ul;
+ xloops *= 4;
+ asm("mull %%edx"
+ :"=d" (xloops), "=&a" (d0)
+ :"1" (xloops), "0" (lpj * (HZ / 4)));
+
+ delay_loop(++xloops);
+}
+EXPORT_SYMBOL(udelay_initialized);
+
void __udelay(unsigned long usecs)
{
__const_udelay(usecs * 0x000010c7); /* 2**32 / 1000000 (rounded up) */
diff --git a/drivers/usb/host/xhci-pci.c b/drivers/usb/host/xhci-pci.c
index 954abfd..b6a7437 100644
--- a/drivers/usb/host/xhci-pci.c
+++ b/drivers/usb/host/xhci-pci.c
@@ -302,6 +302,21 @@ static int xhci_pci_probe(struct pci_dev *dev, const struct pci_device_id *id)
/* USB-2 and USB-3 roothubs initialized, allow runtime pm suspend */
pm_runtime_put_noidle(&dev->dev);
+ do {
+ int count = 1000;
+
+ pr_notice("udelay_test uninitialized ---->start\n");
+ while (count-- > 0)
+ udelay_uninitialized(100);
+ pr_notice("udelay_test uninitialized ---->end\n");
+
+ count = 1000;
+ pr_notice("udelay_test initialized ---->start\n");
+ while (count-- > 0)
+ udelay_initialized(100);
+ pr_notice("udelay_test initialized ---->end\n");
+ } while (0);
+
return 0;
put_usb3_hcd:
diff --git a/include/asm-generic/delay.h b/include/asm-generic/delay.h
index 0f79054..200ab55 100644
--- a/include/asm-generic/delay.h
+++ b/include/asm-generic/delay.h
@@ -9,6 +9,8 @@ extern void __udelay(unsigned long usecs);
extern void __ndelay(unsigned long nsecs);
extern void __const_udelay(unsigned long xloops);
extern void __delay(unsigned long loops);
+extern void udelay_uninitialized(unsigned long xloops);
+extern void udelay_initialized(unsigned long xloops);
/*
* The weird n/20000 thing suppresses a "comparison is always false due to
next prev parent reply other threads:[~2017-01-24 4:44 UTC|newest]
Thread overview: 42+ messages / expand[flat|nested] mbox.gz Atom feed top
2016-11-15 6:02 [PATCH v5 0/4] usb: early: add support for early printk through USB3 debug port Lu Baolu
2016-11-15 6:02 ` [PATCH v5 1/4] usb: dbc: early driver for xhci debug capability Lu Baolu
2017-01-19 9:37 ` Ingo Molnar
2017-01-20 2:47 ` Lu Baolu
2017-01-22 9:04 ` Ingo Molnar
2017-01-24 4:44 ` Lu Baolu [this message]
2017-01-24 8:20 ` Ingo Molnar
2017-01-25 5:28 ` Lu Baolu
2017-01-25 9:23 ` Ingo Molnar
2017-01-25 9:57 ` Peter Zijlstra
2017-01-25 12:27 ` Lu Baolu
2017-01-25 14:38 ` Peter Zijlstra
2017-01-25 15:51 ` Lu Baolu
2017-01-25 16:16 ` Peter Zijlstra
2017-01-26 3:37 ` Lu Baolu
2017-01-26 7:19 ` Ingo Molnar
2017-01-26 7:49 ` Lu Baolu
2017-01-26 8:17 ` Ingo Molnar
2017-01-26 10:28 ` Peter Zijlstra
2017-01-26 16:01 ` Ingo Molnar
2017-01-26 17:39 ` Peter Zijlstra
2017-01-27 6:51 ` Ingo Molnar
2017-02-09 5:59 ` Lu Baolu
2017-01-26 7:22 ` Ingo Molnar
2017-02-09 7:37 ` Lu Baolu
2017-01-25 12:17 ` Lu Baolu
2017-01-26 3:26 ` Lu Baolu
2016-11-15 6:02 ` [PATCH v5 2/4] x86: add support for earlyprintk via USB3 debug port Lu Baolu
2017-01-19 9:38 ` Ingo Molnar
2017-01-20 2:48 ` Lu Baolu
2016-11-15 6:02 ` [PATCH v5 3/4] usb: serial: usb_debug: add support for dbc debug device Lu Baolu
2017-01-19 9:39 ` Ingo Molnar
2017-01-20 2:50 ` Lu Baolu
2016-11-15 6:02 ` [PATCH v5 4/4] usb: doc: add document for USB3 debug port usage Lu Baolu
2017-01-19 9:41 ` Ingo Molnar
2017-01-20 2:53 ` Lu Baolu
2017-01-18 6:20 ` [PATCH v5 0/4] usb: early: add support for early printk through USB3 debug port Lu Baolu
2017-01-19 9:06 ` Greg Kroah-Hartman
2017-01-19 9:09 ` Ingo Molnar
2017-01-19 11:24 ` Mathias Nyman
2017-01-19 9:12 ` Ingo Molnar
2017-01-20 2:56 ` Lu Baolu
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=5886DBB7.4070501@linux.intel.com \
--to=baolu.lu@linux.intel.com \
--cc=gregkh@linuxfoundation.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-usb@vger.kernel.org \
--cc=mathias.nyman@linux.intel.com \
--cc=mingo@kernel.org \
--cc=mingo@redhat.com \
--cc=peterz@infradead.org \
--cc=tglx@linutronix.de \
--cc=x86@kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).